Dear PyMVPA experts,<br><br>I successfully ran through the following code using Python 2.6.2+PyMVPA 2.2.0:<br><br>from mvpa2.tutorial_suite import * <br>ds=get_raw_haxby2001_data(roi='vt') <br>part=NFoldPartitioner()<br>
sl=sphere_gnbsearchlight(GNB(),part,postproc=mean_sample())<br>map=sl(ds)<br><br>However, the same code with a new partitioner led to errors:<br><br>part=ChainNode([NFoldPartitioner(),<br>                Balancer(limit='partitions',apply_selection=True)],<br>
                space='partitions')<br><br>The error messages were:<br><br>Traceback (most recent call last):<br>  File "debug.py", line 9, in <module><br>    map=sl(ds)<br>  File "/usr/local/lib/python2.6/dist-packages/mvpa2/base/learner.py", line 239, in __call__<br>
    return super(Learner, self).__call__(ds)<br>  File "/usr/local/lib/python2.6/dist-packages/mvpa2/base/node.py", line 96, in __call__<br>    result = self._call(ds)<br>  File "/usr/local/lib/python2.6/dist-packages/mvpa2/measures/searchlight.py", line 141, in _call<br>
    results = self._sl_call(dataset, roi_ids, nproc)<br>  File "/usr/local/lib/python2.6/dist-packages/mvpa2/measures/adhocsearchlightbase.py", line 354, in _sl_call<br>    if not (np.all(<a href="http://p.sa">p.sa</a>[targets_sa_name].value == labels)):<br>
ValueError: shape mismatch: objects cannot be broadcast to a single shape<br><br>Any ideas about what's wrong here?<br><br>Thanks!<br>Tren<br>